Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Oct 14.
Published in final edited form as: J Clin Exp Neuropsychol. 2022 Nov 28;44(8):592–603. doi: 10.1080/13803395.2022.2149703

The neurocognitive implications of depression and socioeconomic status in people with HIV

Kaleigh E Fidaleo a,b, Micah J Savin a,c, Maral N Aghvinian a,c, Angela C Summers a,c, Alyssa Arentoft d, Desiree Byrd c,e, Heining Cham a, Monica Rivera Mindt a,c
PMCID: PMC13660556  NIHMSID: NIHMS2211874  PMID: 36440814

Abstract

Objective:

This cross-sectional study investigates the independent and interactive effects of depression and socioeconomic status (SES) on neurocognition in a diverse sample of people with HIV (PWH).

Method:

The sample of 119 PWH (71% Latinx, 27% female) completed comprehensive neurocognitive and psychosocial evaluations and were separated into two groups: those with a history of depression diagnosis (n = 47) and those without (n = 72).

Results:

The results of regression analyses indicated that lifetime depression was not associated with lower SES nor with worse neurocognitive performance on any neurocognitive outcome. However, a significant main effect of SES was observed on the Hopkins Verbal Learning Test (total), indicating that higher SES was associated with better verbal learning performance (B= .11, SE = .05, p< .02). Lastly, the results revealed an interactive effect of lifetime depression and SES, such that individuals with depression and higher SES performed better on tests of attention/working memory (i.e., WAIS-III Letter-Number Sequencing, B= .08, SE = .04, p< .02; Paced Auditory Serial Addition Test, B= .39, SE = .16, p< .02).

Conclusions:

Depression and SES appear to play an important role in the neurocognitive performance of PWH. Specifically, higher SES appears to have a protective effect on attention/working memory among PWH only if they have co-morbid history of lifetime depression.

Keywords: Neurocognition, depression, socioeconomic status, HIV, health disparities

Human immunodeficiency virus (HIV) disproportionately impacts marginalized populations, including those with low socioeconomic status (SES), and is highly co-morbid with both depression and neurocognitive impairment. Depression is the most common neuropsychiatric complication experienced by people with HIV (PWH; Nanni et al., 2015). PWH and depression experience worse clinical outcomes (Kemeny, 2003; Leserman et al., 1997; Page-Shafer et al., 1996; Pellowski et al., 2013), functional outcomes (Heaton et al., 2004a; Marquine et al., 2018; Morgan et al., 2012), and medication adherence (Gonzalez et al., 2011). However, the combined and interactive effects of depression and SES on neurocognitive functioning in PWH remains poorly understood.

Rates of both past and current depression are high in PWH, and both are associated with neurocognitive impairment (Heaton et al., 2015; Rubin & Maki, 2019), including earlier neurocognitive decline (Heaton et al., 2015), faster decline to symptomatic HIV-associated neurocognitive disorder (HAND; Grant et al., 2014), and more notable deficits in psychomotor speed, speed of information processing, attention/working memory, executive functioning, delayed recall, verbal fluency, and global functioning (Bragança & Palha, 2011; Paolillo et al., 2020; Shimizu et al., 2011). However, some studies suggest that PWH and depression are not differentially susceptible to neurocognitive impairment (Armstrong et al., 2017; Cole et al., 2007; Rubin & Maki, 2019). These inconsistent findings may be due to differences in how depression is operationalized across studies (Castellon et al., 2006). For instance, the tendency to measure depression at a single time-point, rather than using a long-term, lifetime approach, may obscure the “cumulative burden” of depression on neurocognitive health (Paolillo et al., 2020)

While depression is a salient variable in this population, other factors such as SES are also pertinent to a comprehensive understanding of HIV and its outcomes. The HIV epidemic is concentrated in impoverished neighborhoods and is overrepresented in locations with wider income gaps and greater stress, including minority stress (Pellowski et al., 2013), less accessible treatment, and greater HIV mortality (Cunningham et al., 2005). Research finds the risk for HIV infection increases as income inequality increases and, further, that HIV diagnosis rates are lower in areas with more non-Latinx white individuals (Gant et al., 2012). However, further research suggests that racial minorities are at a higher risk of contracting HIV not due to their racial/ethnic identity, but because they are more likely to come from a low SES than non-Latinx whites (Denning & DiNenno, 2010).

SES is also independently associated with neurocognitive function among PWH. Adult SES is positively correlated with neurocognitive performance in verbal fluency, processing speed, and executive functioning; further, SES levels are highest among neurocognitively unimpaired individuals and lowest among individuals diagnosed with HIV-associated dementia (HAD; Arentoft et al., 2015). Arentoft et al., 2015 also found that PWH in a higher SES stratum are over 10 times more likely to be neurocognitively unimpaired than their counterparts in a social stratum just one level below. These findings remained significant after controlling for race, illustrating that SES independently accounts for variance in neurocognitive function among PWH and suggesting that even minor differences in SES can have serious repercussions for neurocognitive health. Further, neighborhood-level childhood SES is also associated with neurocognitive function (i.e., impairments in learning/memory and executive functioning) among PWH (Thames et al., 2018). This effect was only observed in PWH, not in HIV-negative controls, suggesting that CNS compromises due to HIV may increase one’s susceptibility to the negative impact of social adversity.

People from low SES backgrounds may also be at a greater risk for depression. This has been observed in the general US population (Bruce, 1991; Lorant et al., 2003) as well as international samples of PWH (Bhatia & Munjal, 2014; Feuillet et al., 2017; Unnikrishnan et al., 2012). For example, among PWH in Portugal, low SES was significantly associated with more severe depression, and more severe depression was significantly associated with worse neurocognitive performance (Bragança & Palha, 2011). However, studies in the USA have not directly examined the relationship between SES and depression in PWH.

No studies to date have simultaneously examined the individual and interactive effects of lifetime depression and SES on neurocognitive impairment in this population, particularly for Latinx PWH. It is possible that SES moderates the association between depression and neurocognitive impairment in PWH and the lack of accounting for SES in such prior studies may help explain some of the equivocal findings in the depression and HAND literature to date. Further understanding these variables and their combined influence on PWH might allow for more rigorous and tailored psychiatric assessment, treatment, and prevention efforts and therefore this research gap is an impediment to individual and public health.

The current study aimed to address this gap in the literature by investigating: 1) how the rate of lifetime depression differed amongst PWH of varying SES backgrounds; 2) the individual effects of lifetime depression and SES on neurocognitive functioning; and 3) the interactive effects of lifetime depression and SES on neurocognitive functioning. The current study included a diverse sample of Latinx and non-Latinx white PWH in investigating these aims. We hypothesized that 1) PWH with lower SES would show higher rates of lifetime depression than those of a higher SES; 2) compared to people without lifetime depression, those with lifetime depression would be more likely to demonstrate worse neurocognitive functioning (i.e., motor functioning, processing speed, attention/working memory, executive functioning, memory, and verbal fluency); 3) lower SES would be associated with worse neurocognitive functioning (i.e., learning, memory, verbal fluency, processing speed, and executive functioning); and 4) SES and depression would significantly interact to moderate neurocognitive performance. Specifically, it was expected that SES would moderate the association between depression and neurocognitive performance such that the level of neurocognitive functioning of those with a higher SES would remain mostly consistent regardless of lifetime depression diagnosis, whereas those with a lower SES would show lower levels of neurocognitive functioning if they have a lifetime depression diagnosis compared to if they do not have a lifetime depression diagnosis.

Methods

Participants

This was a cross-sectional retrospective study based on archival data from an NIH-funded parent study: the Medication Adherence Study (MÁS), which utilized a community-based participatory research approach to examine health inequities in medication adherence among HIV-positive Latinx and non-Latinx white adults (PI: M. Rivera Mindt; K23MH078718). The current study included an ethnoculturally diverse sample of 119 PWH who were assessed at the Icahn School of Medicine at Mount Sinai (ISMMS) in East Harlem, New York. Due to the plethora of research focused on depression and neurocognition, and the dearth of research focused on SES and neurocognition, the sample was delineated into two groups: participants with lifetime (current and/or history of) Major Depressive Disorder based on a CIDI diagnosis (i.e., MDD+; n = 47) and a comparison group comprised of participants with no lifetime MDD diagnoses (i.e., MDD−; n = 72). Within the MDD+ group, all 47 participants had a history of MDD, and 20 participants (43%) were experiencing a current depressive episode.

Inclusion criteria

The parent study adhered to the following inclusion criteria: participants needed to be between 18–80 years old, provide documentation of their HIV seropositive status, be fluent in English, be taking at least one antiretroviral therapy medication in pill form for a minimum of 12 weeks, and be capable of taking their medication without assistance. Further, to demonstrate capacity to provide informed consent, participants were asked to explain the study and participant expectations in their own words before signing the MÁS consent form. To be included in the current study, participants were required to have completed the Composite International Diagnostic Interview 2.1 (CIDI), the Hollingshead Four Factor Scale of Social Status, and the neurocognitive battery. Forty-four participants from the parent study were missing this data and therefore were not included in the final sample of 119 participants. These participants showed no significant differences in gender, race, age, education, or global cognition (all ps > .05; via independent samples t-tests and χ2 tests of association).

Exclusion criteria

Participants were excluded from the parent study if they had injuries, illnesses, or surgeries capable of confounding neurocognitive functioning, including: brain surgery; a head injury with loss of consciousness for longer than 12 hours; a penetrating head injury; a diagnosis of schizophrenia, schizoaffective disorder, or bipolar disorder, unstable or active psychosis; or a diagnosis of epilepsy, seizure disorder, Parkinson’s Disease, Lupus, end-stage renal disease requiring dialysis, Multiple Sclerosis, stroke, brain cancer, or a brain tumor. Individuals who received their HIV medication in pre-sorted compliance packaging and those intending to change their HIV medication within one month of study enrollment were also excluded. As mentioned previously, participants missing data from the CIDI or data necessary to calculate the Hollingshead Four Factor Scale of Social Status were excluded from analysis.

Procedures

After providing informed written consent, participants completed a neuromedical and laboratory evaluation, a comprehensive neurocognitive evaluation, and a psychosocial interview involving assessments of psychiatric, socioeconomic, and demographic information. All study procedures were approved by the Institutional Review Board at ISMMS.

Materials and measures

Neurocognitive evaluation

Participants completed a comprehensive neurocognitive battery which assessed the following domains: learning, memory, processing speed, attention/working memory, executive functioning, verbal fluency, and motor functioning. All neurocognitive evaluations were administered and scored by research assistants supervised by board-certified clinical neuropsychologists. This battery is well-validated for the evaluation of neurocognition in PWH (Antinori et al., 2007; Heaton et al., 2010; Justice et al., 2004). Table 1 provides a list of the specific tests included in the battery.

Table 1.

Neurocognitive test battery and corresponding normative data.

Neurocognitive Domain and Tests
Learning
 Hopkins Verbal Learning Test – Revised Benedict et al., 1998 1
 Brief Visuospatial Memory Test – Revised Benedict, 1997 1
Memory/Delayed Recall
 Hopkins Verbal Learning Test – Revised Benedict et al., 1998 1
 Brief Visuospatial Memory Test – Revised Benedict, 1997 1
Processing Speed
 WAIS-III Digit Symbol Heaton et al., 2003 1,2,3,4
 WAIS-III Symbol Search Heaton et al., 2003 1,2,3,4
 Trail Making Test (Part A) (Heaton et al., 2004b)1,2,3,4
Attention/Working Memory
 WAIS-III Letter-Number Sequencing Heaton et al., 2003 1,2,3,4
 PASAT Total Correct (Heaton et al., 2004b)1,2,3,4
Executive Functioning
 Wisconsin Card Sorting Task-64 Item Version Kongs et al., 2000 1,2
 Trail Making Test (Part B) (Heaton et al., 2004b)1,2,3,4
Verbal Fluency
 Controlled Oral Word Association Test (FAS) (Heaton et al., 2004b)1,2,3,4
 Semantic (Animal) Fluency (Heaton et al., 2004b)1,2,3,4
Motor Functioning
 Grooved Pegboard Time (Heaton et al., 2004b)1,2,3,4
Open in a new tab

WAIS = Wechsler Adult Intelligence Scales; PASAT = Paced Auditory Serial Addition Test.

Test. Normative data is demographically corrected for age (1),

education (2),

sex (3),

ethnicity (4)

Raw test scores from each neurocognitive domain were used to evaluate study hypotheses. This was done in lieu of adjusting scores for demographic variables, as adjusting for education might mitigate observed relationships between SES and neurocognitive performance. Instead, significant relationships between neurocognitive performance and demographic variables were controlled by including demographic variables as covariates in our statistical models (see Results section). The following tests were used to evaluate performance in seven neurocognitive domains:

  • Verbal fluency: Controlled Oral Word Association F-A-S total (COWAT: Phonemic Fluency), Semantic (Animal) Fluency total

  • Executive functioning: Trail Making Test (TMT) Part B time, Wisconsin Card Sorting Test-64 (WCST) total perseverative responses

  • Processing speed: Trail Making Test (TMT) Part A time, Wechsler Adult Intelligence Scale-III (WAIS-III) Digit Symbol total, WAIS-III Symbol Search total

  • Attention/working memory: WAIS-III Letter-Number Sequencing total (LNS), Paced Auditory Serial Addition Test total (PASAT)

  • Learning: Hopkins Verbal Learning Test total (HVLT total), Benton Visual Memory Test total (BVMT total)

  • Memory: Hopkins Verbal Learning Test delay (HVLT delay), Benton Visual Memory Test delay (BVMT delay)

  • Motor: Grooved Pegboard dominant hand time, and Grooved Pegboard nondominant hand time

Demographically corrected t-scores were only used to assign clinical ratings. Interpretation of neurocognitive test performance was based upon ratings reported in (Woods et al., 2004).

Psychosocial evaluation

Participants completed the CIDI to evaluate current and lifetime psychiatric diagnoses. Participants also completed a socioeconomic and demographic questionnaire to gather the information necessary to estimate SES using the Hollingshead Four Factor Scale of Social Status.

Composite International Diagnostic Interview 2.1 (CIDI)

The CIDI version 2.1 assesses psychiatric disorders across the lifespan with close adherence to definitions and criteria of the DSM-IV (American Psychiatric Association, 1994) and ICD-10 (World Health Organization, 2016). The range of diagnostic coverage implemented in the current study included substance use disorders, schizophrenic disorders, and mood disorders, however the full CIDI 2.1 measures various additional classes of psychiatric disorders. The interview is comprised of up to 288 questions and can last from less than an hour to 2–3 hours (Wittchen et al., 1991). Question content includes inquiries regarding symptoms, specifically symptom presence, severity, recency, frequency, and duration (Wittchen et al., 1991). The CIDI has excellent interrater reliability in all diagnostic sections (agreement = 97–99%; κ’s = .67-.99; Wittchen et al., 1991). It has also been found to be appropriate for use across various settings, countries, and cultures (Wittchen et al., 1991). Any participants in the current study with at least one current and/or past episode of major depression, as defined by the CIDI, were categorized as having lifetime depression.

Socioeconomic and demographic evaluation

The Hollingshead Four Factor Scale of Social Status, specifically the Adult Hollingshead Index of Social Position (ISP), was used as a proxy for estimating participant SES as it is one of the most broadly cited measures of SES in the literature (Cirino et al., 2002). The Hollingshead scale has demonstrated strong interrater reliability (agreement = 89%; r = .91; κ = .68), adequate concurrent validity with other measures of SES (Nakao and Treas Socioeconomic Index of Occupation and Blishen, Carrol, and Moore Socioeconomic Index for Occupations in Canada; r’s = .81 – .86), and modest convergent validity with the Kaufman Brief Intelligence Test composite IQ score (r= .43; Cirino et al., 2002).

To calculate Adult Hollingshead ISP scores, demographic and sociocultural information was gathered via a brief interview. Education was measured as the highest level of formal education. Participants reported their most recent, longest held, and highest-level jobs, and this occupational history was coded using methods provided by Hollingshead (2011). Participants’ highest-ranked job (in correspondence with Hollingshead ratings) was then used to compute their Hollingshead ISP score. This score was estimated using the Hollingshead scale and the following equation: (occupation scale value * 5) + (education scale value * 3; Hollingshead, 2011). For context, occupational scores are ranked from 1 (e.g., “farm laborers and menial service workers”) to 9 (e.g., “higher executives, proprietors of large businesses, and major professionals”). Scores range from 8 to 66, and fall into 5 social strata (i.e., < 20 = strata 1, 20–29 = strata 2, 30–39 = strata 3, 40–54 = strata 4, 55+ = strata 5). Higher scores correspond to higher SES. The procedure and scoring are described in greater detail in (Arentoft et al., 2015).

Neuromedical and laboratory evaluation

The neuromedical evaluation briefly assessed participants’ HIV and neurological (personal and familial) history. Participants provided HIV-related information such as Nadir CD4 T-cell count, date of HIV diagnosis, and HIV treatment history. Neurological history included personal history of brain injury/surgery and family history of neurological diseases such as Alzheimer’s Disease.

Blood samples were drawn to evaluate current absolute CD4 T-cell count and HIV RNA plasma viral load. A urine toxicology test was performed to assess for drug use that might interfere with neurocognitive performance (i.e., use of amphetamines, barbiturates, benzo-diazepines, tricyclic antidepressants, opiates, phencyclidine, methadone, methamphetamine, cocaine, and cannabinoids).

Statistical analysis

Results were analyzed using SPSS Version 25 (IBM Corp, 2017) and R 4.2.0 (R Core Team, 2022). Before any descriptive or inferential analyses were conducted, all variables of interest were evaluated for normality. Analysis of histograms, Q-Q plots, and measures of skewness and kurtosis indicated that all pertinent variables met criteria for normality, allowing for the utilization of parametric statistical tests. If regression diagnostics showed non-normality and heteroskedasti-city of residuals, we used Box-Cox transformations on the dependent variable (Venables & Ripley, 2002). The following dependent variables necessitated transformation: WCST total perseverative responses, TMT Part A time, TMT Part B time, Grooved Pegboard dominant hand time, and Grooved Pegboard nondominant hand time.

To evaluate potential confounds, the MDD diagnostic groups (e.g., MDD+ vs. MDD−) were assessed for group differences on demographic, psychosocial, and clinical variables (i.e., age, gender, ethnicity, education, SES, substance use [based on CIDI diagnoses and urine toxicology results], HIV duration, HIV viral load, CD4 T-cell count, and nadir CD4 T-cell count) using one-way ANOVAs and χ2 tests of association. The groups did not significantly differ on any of the aforementioned variables (all ps > .10). Inferential statistical analyses, implemented to assess study hypotheses, followed. Independent samples t-tests were conducted to evaluate group differences in SES. To evaluate group differences in neurocognitive performance, associations between SES and neurocognitive performance, and to determine if SES moderated the relationship between MDD and neurocognitive functioning, linear regression analyses were conducted in which MDD, SES, and the interaction term of these two variables served as the independent variables and raw neurocognitive test scores served as the dependent variables (see, Table 1 for list of included neurocognitive tests).

Results

The sample was predominantly male (73%), largely Latinx (71%), and had a mean age of 47.1 (SD = 8.5). Latinx participants were recruited from any/all racial backgrounds and all remaining participants identified as non-Latinx white. Participants completed an average of 12.9 (SD = 2.7) years of education. HIV viral load (HIV RNA log10 cutoff: > 1.69) was detectable in 34% of the sample and the viral load ranged from 19–1,114,637 with a median of 24. Further, only 11% of participants had a CD4 count below 200. The MDD+ and MDD− groups did not significantly differ on any of the aforementioned variables. Demographic and clinical information is summarized in Table 2.

Table 2.

Demographics and clinical characteristics of the sample (N = 119).

Entire Sample (N= 119) M (SD) or % MDD+ (n = 47) M (SD) or % MDD− (n = 72) M (SD) or % F
Demographics
 Age (in years) 47.1 (8.5) 47.8 (8.5) 46.6 (8.6) .55
 % Latinx 70.6% 66.0% 73.6% .79
 Education (in years) 12.9 (2.7) 13.1 (2.8) 12.8 (2.7) .34
 % Male 73.1% 72.3% 73.6% .02
 SES (Hollingshead)
  Score 38.3 (13.9) 40.9 (13.1) 36.6 (14.2) 1.10
  Strata 3 4 3
Clinical Characteristics
 % CD4 Count <200 10.9% 15.2%  9.0% 1.04
 % Detectable HIV 33.6% 40.0% 32.8%  .59
 Viral Load (log10 > 1.69)
Open in a new tab

MDD = Major Depressive Disorder.

On average, the MDD+ group met criteria for strata 4 on the Hollingshead scale (M= 40.91, SD = 13.11) and the MDD− group met criteria for strata 3 (M= 36.56, SD = 14.18) (see, Figure 1). Concerning neurocognition, the MDD+ group performed in the low average range on the domains of learning (M = 40.22, SD = 12.27), memory (M = 40.70, SD = 12.52), attention/working memory (M = 44.72, SD = 9.11), executive functioning (M = 44.03, SD = 10.81), and motor functioning (M = 42.15, SD = 9.39) and the average range for processing speed (M = 47.65, SD = 9.35) and fluency (M = 45.23, SD = 11.72; Woods et al., 2004). The MDD− group performed in the definite mild impairment range on the domains of learning (M = 36.17, SD = 11.42) and memory (M = 37.24, SD = 12.06), the low average range on domains of attention/working memory (M = 41.93, SD = 8.09), executive functioning (M = 43.61, SD = 10.77), and motor functioning (M = 41.64, SD = 10.48), and the average range on domains of processing speed (M = 46.24, SD = 8.51) and fluency (M = 45.25, SD = 10.51; Woods et al., 2004).

Figure 1.

Figure 1.

Open in a new tab

Between group differences in adult Hollingshead ISP scores. Note. MDD = Major Depressive Disorder, † .05 ≥ p < .10

To test the first study hypothesis, an independent samples t-test was conducted to evaluate group differences in SES, as measured by Hollingshead ISP scores. This analysis indicated a trend-level between-group difference in SES, t(117) = −1.69, p< .09, d= .32. The MDD + group (M= 40.91, SD = 13.11) scored higher on the Hollingshead scale than the MDD− group (M= 36.56, SD = 14.18). These results are depicted in Figure 1. Because higher Hollingshead scores are indicative of higher SES, the first study hypothesis was not supported.

Table 3 summarizes a series of linear regression analyses that were conducted to test second, third, and fourth hypotheses. The predictor variables in the model included MDD status (0 for “MDD−” and 1 for “MDD+”), Adult Hollingshead ISP scores (scores ranging from 8–66), and the interaction term for MDD status x Adult Hollingshead ISP. Because raw scores were used instead of demographically adjusted T-scores, demographic variables (e.g., ethnicity, gender, age, and education) were included as covariates. Outcome variables included raw neurocognitive test scores, with a test of linear regression conducted for each test score.

Table 3.

Linear regressions predicting neurocognitive functioning by MDD grouping, adult Hollingshead ISP, and their interaction among the full sample (N= 119).

Outcome R 2 Δ R2 MDD Status
 Adult Hollingshead ISP
 MDD Grouping x Hollingshead ISP
B SE B B SE B B SE B
COWAT .15** .021 −11.53† 6.90 −.02 .13 .27 .17
SF (Animal) .11  .066 −9.18 3.30 −.02 .06 .21 .08
Digit Sym .16** .003  5.91 8.41  .27† .15 −.12 .20
Sym Search .18** .005 −1.66 4.22 .10 .08  .08 .10
LNS .15*  .043 −2.39 1.53 .01 .03   .08* .04
PASAT .15*  .047 −13.58* 6.84 .01 .13  .39* .16
HVLT total  .27*** .006 −1.64 2.57  .11* .05 .06 .06
BVMT total .13*  .006 5.87 4.22  .003 .08 −.09 .10
HVLT delay  .23*** .005 −.67 1.24 .05†  .02 .02 .03
BVMT delay .17** .009 2.44 1.62 .01  .03 −.04 .04
TMT B .09  .017  .07  .05 >.001  .001  −.002  .001
WCSTPR .05  >.001  −.28  .41 −.006  .008  .002 .01
TMT A .06  >.001  −.03  .08 −.001  .001 >.001  .002
Peg DH .10  .002 >.001   >.001 >.001  >.001  >.001 >.001
Peg NDH .10  .002 >.001   >.001 >.001  >.001  >.001 >.001
Open in a new tab

MDD = Major Depressive Disorder, COWAT = Controlled Oral Word Association Test (FAS), SF = Semantic Fluency, Digit Sym = WAIS-III Digit Symbol, Sym Search = WAIS-III Symbol Search, LNS = WAIS-III Letter-Number Sequencing, PASAT = Paced Auditory Serial Arithmetic, HVLT = Hopkins Verbal Learning Test, BVMT = Brief Visuospatial Memory Test, TMT B = Trails Making Test (Part B), WCSTPR = Wisconsin Card Sorting Task64 Perseverative Responses, TMT A = Trails Making Test (Part A), Peg DH = Grooved Peg Board dominant hand time, Peg NDH = Grooved Ped Board nondominant hand time. Predictors variable is Lifetime MDD Grouping, moderating variable is Adult Hollingshead ISP Score. Predictors variable is Lifetime MDD Grouping, moderating variable is Adult Hollingshead ISP Score.

†

.05 < p < .10

*

p< .05

**

p< .01

***

p< .001

The results of this series of regression analyses indicated that our model accounted for a significant amount of variance (13%-27%) in the following tests: COWAT: Phonemic Fluency, WAIS-III Digit Symbol, WAIS-III Symbol Search, WAIS-III LNS, PASAT, HVLT total, BVMT total, HVLT delay, and BVMT delay.

To evaluate our second hypothesis, we assessed MDD group differences in neurocognitive performance by observing the main effects of MDD on neurocognitive performance in our linear regression models. Results showed a significant main effect of MDD on the PASAT, indicating that the MDD− group scored lower than the MDD+ group by 13.6 points. Because a lower score on the PASAT is indicative of worse performance, this hypothesis was not supported.

To evaluate our third hypothesis, we assessed the impact of SES on neurocognitive performance by observing the main effects of SES on neurocognitive performance in our linear regression models. A significant main effect of SES was observed on the HVLT total, indicating that this hypothesis was partially supported.

However, while each of the aforementioned omnibus models were significant, the interaction term (representing the combined variable of MDD grouping and Adult Hollingshead ISP) only significantly accounted for a significant amount of variance in tests of attention/working memory, WAIS-III LNS (illustrated in Figure 2) and PASAT (illustrated in Figure 3). In the remaining models, only covariates significantly predicted variance in the outcome variables. Thus, the fourth hypothesis was partially supported.

Figure 2.

Figure 2.

Open in a new tab

Interaction of MDD and SES on WAIS-III LNS performance. Note. MDD = Major Depressive Disorder, SES = socioeconomic status, LNS = Letter-Number Sequencing

Figure 3.

Figure 3.

Open in a new tab

Interaction of MDD and SES on PASAT performance. Note. MDD = Major Depressive Disorder, SES = socioeconomic status, PASAT = Paced Auditory Serial Addition Test

Discussion

The current study aimed to investigate the individual and interactive effects of depression and socioeconomic status, an important social determinant of health, on neurocognitive functioning in an ethnoculturally diverse sample of PWH. Further, the present study sought to determine if PWH with lower SES experience higher rates of depression than those with higher SES. Contrary to hypotheses, results indicated that participants with depression were not more likely to have a lower SES than participants without depression, nor were the participants with depression more neurocognitively impaired than participants without depression. However, consistent with hypotheses, higher SES was associated with stronger neurocognitive functioning, and depression and SES interacted to predict neurocognitive performance on tests of attention/working memory.

Contrary to study hypotheses, as well as evidence from past research (Lorant et al., 2003; Miech & Shanahan, 2000), lower current SES was not significantly associated with a diagnosis of depression in PWH. This unexpected finding might be related to gender. Higher rates of depression and a stronger association between SES and depression have been reported among women (Gilman et al., 2002; Nanni et al., 2015). Cuellar and Roberts (1997) found that while low SES and female gender were the greatest risk factors for depression in a Latinx sample without HIV, gender accounted for more variance in depression scores than SES. It is therefore possible that the current study may have failed to replicate these findings due to our predominantly male sample. Other factors associated with depression in PWH, such as drug use, poor self-efficacy, and lack of social support (Nanni et al., 2015), were not examined in the current study and could potentially account for the lack of results.

While previous research on the association between depression and neurocognitive performance has yielded inconsistent findings, the current study predicted that lifetime depression would be significantly associated with poorer neurocognitive functioning. This hypothesis was not supported; for most neurocognitive outcomes, no significant main effect of depression was observed. A significant main effect of depression on neurocognitive performance was only found for the PASAT, a measurement of attention and working memory. Contrary to expectations, and previous research (Bragança & Palha, 2011), the MDD− group demonstrated more impairment than the MDD+ group. It is possible that the use of a binary diagnosis (i.e., MDD+ vs. no MDD−), rather than dimensional scores, inadvertently obfuscated the relationships between neurocognition and relevant components of depression (Castellon et al., 2006). Another potential explanation is that the impact of HIV on neurocognition is strong enough to mask the influence of depression.

Consistent with hypotheses, higher current SES was associated with better neurocognitive functioning, albeit only in the HVLT total score, a measurement of verbal learning. This finding supports previous research that found lower childhood SES was correlated with increased risk for impairments in learning/memory and executive functioning in PWH (Thames et al., 2018). This work in conjunction with the results of the current study suggests that vulnerability to neurocognitive impairment in association with SES is present early on and as well as throughout the lifespan.

Additionally, as hypothesized, SES and depression significantly interacted to predict neurocognitive performance in the sample, although not as expected. The interactive effects of SES and depression were significantly associated with tests of attention/working memory such that PWH and lifetime depression performed better as SES increased. Interestingly, SES did not influence performance in attention/working memory among those without depression. Therefore, the results suggest that amongst depressed, ethnoculturally diverse PWH, neurocognitive performance in attention/working memory is at least partly contingent on SES.

The neural substrates of the related, yet distinct, domains of attention and working memory maintain various distinctions, however both domains utilize the right prefrontal cortex (Mayer et al., 2007). Depression has been associated with significant reductions in levels of early gene expression in the medial prefrontal cortex, an indication of deficient neuronal activity (Covington et al., 2010). Because the prefrontal cortex is involved in attention, working memory, and depression, future research should further explore whether this area may be more vulnerable to deficits in low SES PWH with a history of depression than their high SES counterparts. Higher SES could be a protective factor against attention/working memory impairment in PWH and depression. Research suggests that those with higher SES may avoid some risk factors and benefit from some protective factors (e.g., greater financial stability, better access to quality healthcare treatment options, more information about maintaining overall health, higher cognitive reserve) that promote healthy neurocognition (Beltran-Najera et al., 2022; Chu & Selwyn, 2008; Liu & Li, 2020; Wood et al., 2002; Zhu et al., 2018).

While the findings of the current study are promising, limitations should also be acknowledged. The current study utilized total depression scores and did not differentiate depression by severity, therefore potentially obscuring nuances in the impact of depression on neurocognition. We did not have information on MDD episode recency in participants without current MDD or detailed MDD-related treatment information. In addition, our measure of SES, the Hollingshead scale, is outdated and occasionally unclear regarding its occupation classification system (Haug & Sussman, 1971). However, previous research that compared multiple measures of SES still recognized the Hollingshead as the most sensitive (Arentoft, 2012). We also used a cross-sectional design, which limited our ability to establish temporal precedence. Further, the restricted range in SES scores among the sample likely compromised our ability to measure how SES predicts differences in depression and neurocognition. This restriction might be explained by the fact that PWH are frequently concentrated in similar social groups; including a sample of PWH with greater range and variability may have mitigated this limitation. This study also focused on a Latinx population, therefore limiting generalizability to the greater population of PWH. Last, evaluating neurocognitive performance by individual test required a larger number of analyses than domain-level comparisons, which increases the risk for Type 1 error.

Future studies can build from the limitations of the current study by involving more specified depression groupings and including a larger sample with more socioeconomic and gender diversity. A larger sample will allow for more statistical power and a more reasonable implementation of Type 1 error and false discovery rate corrections. Longitudinal research is needed to compare the impact of lifelong versus single-episodes of depression as well as explore any differences between major depressive episodes, periods of remission, and persistent depressive symptomology. Further, as recommended by Castellon et al. (2006), future efforts might examine depression as a dimensional construct rather than singular, cohesive variable. Lastly, additional efforts should be made to clarify the protective value of SES on neurocognition (specifically attention/working memory) in those with and without a history of depression.

The current study represents a first step toward a better understanding of the individual and interactive effects of depression and SES on neurocognition. This area of research will contribute clinically valuable information to the literature that could support more rigorous and socioculturally-tailored psychiatric assessment and intervention to improve individual and public health outcomes, particularly among PWH from underserved and under-researched communities.

Acknowledgments

The authors would like to thank all participants for their contributions to the research and the Harlem Community and Academic Partnership for their support of this work in the community. This research was supported by a K23 from NIMH (K23MH079718) and an R01 from the NIH (R01AG065110-01A1). For additional publications from this grant, please see url: https://www.ncbi.nlm.nih.gov/pmc/?term=K23MH079718

Funding

This work was supported by the National Institute of Mental Health; National Institutes of Health

Footnotes

Disclosure statement

No potential conflict of interest was reported by the author(s).

References

  1. American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). American Psychiatric Publishing, Inc. [Google Scholar]
  2. Antinori A, Arendt G, Becker JT, Brew BJ, Byrd DA, Cherner M, Clifford DB, Cinque P, Epstein LG, Goodkin K, Gisslen M, Grant I, Heaton RK, Joseph J, Marder K, Marra CM, McArthur JC, Nunn M, Price RW, … Wojna VE. (2007). Updated research nosology for HIV-associated neurocognitive disorders. Neurology, 69(18), 1789–1799. 10.1212/01.WNL.0000287431.88658.8b [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Arentoft A. (2012). Socioeconomic predictors of neuropsychological performance in an urban HIV+ cohort. (Publication No.AAI3495848) [Doctoral dissertation, Fordham University]. ETD Collection for Fordham University. [Google Scholar]
  4. Arentoft A, Byrd D, Monzones J, Coulehan K, Fuentes A, Rosario A, Miranda C, Morgello S, & Rivera Mindt M. (2015). Socioeconomic status and neuropsychological functioning: Associations in an ethnically diverse HIV+ cohort. The Clinical Neuropsychologist, 29(2), 232–254. 10.1080/13854046.2015.1029974 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Armstrong NM, Surkan PJ, Treisman GJ, Sacktor NC, Irwin MR, Teplin LA, Stall R, Martin EM, Becker JT, Munro C, Levine AJ, Jacobson LP, & Abraham AG. (2017). Association of long-term patterns of depressive symptoms and attention/executive function among older men with and without human immunodeficiency virus. Journal of Neurovirology, 23(4), 558–567. 10.1007/s13365-017-0527-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Beltran-Najera I, Thompson JL, Matchanova A, Sullivan KL, Babicz MA, & Woods SP. (2022). Neurocognitive performance differences between black and white individuals with HIV disease are mediated by health literacy. The Clinical Neuropsychologist, 36(2), 414–430. 10.1080/13854046.2021.1953147 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Benedict R. (1997). Brief visuospatial meory test–revised professional manual. Psychological Assessment Resources, Inc. [Google Scholar]
  8. Benedict RHB, Schretlen D, Groninger L, & Brandt J. (1998). Hopkins verbal learning test–revised: Normative data and analysis of inter-form and test-retest reliability. The Clinical Neuropsychologist, 12, 43–55. 10.1076/clin.12.1.43.1726 [DOI] [Google Scholar]
  9. Bhatia MS, & Munjal S. (2014). Prevalence of depression in people living with HIV/AIDS undergoing ART and factors associated with it. Journal of Clinical and Diagnostic Research : JCDR, 8(10), WC01–WC04. 10.7860/JCDR/2014/7725.4927 [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Bragança M, & Palha A. (2011). Depression and neurocognitive performance in Portuguese patients infected with HIV. AIDS and Behavior, 15(8), 1879–1887. 10.1007/s10461-011-9973-3 [DOI] [PubMed] [Google Scholar]
  11. Bruce ML. (1991). Poverty and psychiatric status: Longitudinal evidence from the new haven epidemiologic catchment area study. Archives of General Psychiatry, 48(5), 470. 10.1001/archpsyc.1991.01810290082015 [DOI] [PubMed] [Google Scholar]
  12. Castellon SA, Hardy DJ, Hinkin CH, Satz P, Stenquist PK, van Gorp WG, Myers HF, & Moore L. (2006). Components of depression in HIV-1 infection: Their differential relationship to neurocognitive performance. Journal of Clinical and Experimental Neuropsychology, 28(3), 420–437. 10.1080/13803390590935444 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Chu C, & Selwyn PA. (2008). Current health disparities in HIV/AIDS. The AIDS Reader, 18(3), 144–6, 152–8, C3. https://www.patientcareonline.com/view/current-health-disparities-hivaids [PubMed] [Google Scholar]
  14. Cirino P, Sevcik R, Wolf M, Lovett M, & Morris R. (2002). Measuring socioeconomic status: Reliability and preliminary validity of different approaches. Assessment, 9 (2), 145–155. 10.1177/10791102009002005 [DOI] [PubMed] [Google Scholar]
  15. Cole MA, Castellon SA, Perkins AC, Ureno OS, Robinet MB, Reinhard MJ, Barclay TR, & Hinkin CH. (2007). Relationship between psychiatric status and frontal–Subcortical systems in HIV-infected individuals. Journal of the International Neuropsychological Society : JINS, 13(3), 549–554. 10.1017/S135561770707066X [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Covington HE, Lobo MK, Maze I, Vialou V, Hyman JM, Zaman S, LaPlant Q, Mouzon E, Ghose S, Tamminga CA, Neve RL, Deisseroth K, & Nestler EJ. (2010). Antidepressant effect of optogenetic stimulation of the medial prefrontal cortex. Journal of Neuroscience, 30(48), 16082–16090. 10.1523/JNEUROSCI.1731-10.2010 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Cuellar I, & Roberts RE. (1997). Relations of depression, acculturation, and socioeconomic status in a Latino sample. Hispanic Journal of Behavioral Sciences, 19(2), 230–238. 10.1177/07399863970192009 [DOI] [Google Scholar]
  18. Cunningham WE, Hays RD, Duan N, Andersen R, Nakazono TT, Bozzette SA, & Shapiro MF. (2005). The effect of socioeconomic status on the survival of people receiving care for HIV infection in the United States. Journal of Health Care for the Poor and Underserved, 16(4), 655–676. 10.1353/hpu.2005.0093 [DOI] [PubMed] [Google Scholar]
  19. Denning P, & DiNenno E. (2010) Communities in crisis: Is there ageneralized HIV epidemic in impoverished urban areas of the United States? Centers for disease control and prevention. In XVIII international AIDS conference. https://www.cdc.gov/hiv/group/poverty.html [Google Scholar]
  20. Feuillet P, Lert F, Tron L, Aubriere C, Spire B, & Dray-Spira R, & the Agence Nationale de Recherche sur le Sida et les Hépatites Virales (ANRS)-VIH: Enquête sur les personnes atteintes (Vespa2) Study Group. (2017). Prevalence of and factors associated with depression among people living with HIV in France. HIV Medicine, 18 (6), 383–394. 10.1111/hiv.12438 [DOI] [PubMed] [Google Scholar]
  21. Gant Z, Lomotey M, Hall HI, Hu X, Guo X, & Song R. (2012). A county-level examination of the relationship between HIV and social determinants of health: 40 states, 2006-2008. The Open AIDS Journal, 6(1), 1–7. 10.2174/1874613601206010001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Gilman SE, Kawachi I, Fitzmaurice GM, & Buka SL. (2002). Socioeconomic status in childhood and the lifetime risk of major depression. International Journal of Epidemiology, 31(2), 359–367. 10.1093/ije/31.2.359 [DOI] [PubMed] [Google Scholar]
  23. Gonzalez JS, Batchelder AW, Psaros C, & Safren SA. (2011). Depression and HIV/AIDS treatment nonadherence: A review and meta-analysis. Journal of Acquired Immune Deficiency Syndromes, (1999), 58(2. 10.1097/QAI.0b013e31822d490a [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Grant I, Franklin DR, Deutsch R, Woods SP, Vaida F, Ellis RJ, Letendre SL, Marcotte TD, Atkinson JH, Collier AC, Marra CM, Clifford DB, Gelman BB, McArthur JC, Morgello S, Simpson DM, McCutchan JA, Abramson I, Gamst A, … Heaton RK. (2014). Asymptomatic HIV-associated neurocognitive impairment increases risk for symptomatic decline. Neurology, 82(23), 2055–2062. 10.1212/WNL.0000000000000492 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Haug MR, & Sussman MB. (1971). The indiscriminate state of social class measurement. Social Forces, 49(4), 549–563. 10.2307/2576736 [DOI] [Google Scholar]
  26. Heaton RK, Clifford DB, Franklin DR, Woods SP, Ake C, Vaida F, Ellis RJ, Letendre SI, Marcotte TD, Atkinson JH, Rivera Mindt M, Vigil OR, Taylor MJ, Collier AC, Marra CM, Gelman BB, McArthur JC, Morgello S, Simpson DM, … Grant I. (2010). HIV-associated neurocognitive disorders persist in the era of potent antiretroviral therapy. Neurology, 75, 2087–2096. 10.1212/WNL.0b013e318200d72723 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Heaton RK, Franklin DR, Deutsch R, Letendre S, Ellis RJ, Casaletto K, Marquine MJ, Woods SP, Vaida F, Atkinson JH, Marcotte TD, McCutchan JA, Collier AC, Marra CM, Clifford DB, Gelman BB, Sacktor N, Morgello S, Simpson DM, … Teshome M. (2015). Neurocognitive change in the era of HIV combination antiretroviral therapy: The longitudinal CHARTER study. Clinical Infectious Diseases, 60(3), 473–480. 10.1093/cid/ciu862 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Heaton RK, Marcotte TD, Rivera Mindt M, Sadek J, Moore DJ, Bentley H, McCutchan JA, Reicks C, & Grant I, & The HNRC Group. (2004a). The impact of HIV-associated neuropsychological impairment on everyday functioning. Journal of the International Neuropsychological Society, 10 (3), 317–331. 10.1017/s1355617704102130 [DOI] [PubMed] [Google Scholar]
  29. Heaton RK, Miller SW, Taylor MJ, & Grant I. (2004b). Revised comprehensive norms for an expanded Halstead-reitan battery: Demographically adjusted neuropsychological norms for Non-Hispanic Black/African American and Caucasian adults (HRB). Psychological Assessment Resources, Inc. [Google Scholar]
  30. Heaton RK, Taylor MJ, & Manly J. (2003). Demographic effects and use of demographically corrected norms with the WAIS-III and WMS-III. In Tulsky D, Saklofske D, & Chelune G, et al. (Eds.), Clinical Interpretation of the WAIS-III and WMS-III (pp. 181–209). Orlando, FL: Elsevier Science. [Google Scholar]
  31. Hollingshead AB. (2011). Four factor index of social status. Yale Journal of Sociology, 8, 21–52. https://sociology.yale.edu/sites/default/files/files/yjs_fall_2011.pdf#page=21 [Google Scholar]
  32. IBM Corp. (2017). IBM SPSS statistics for windows, version 25.0.
  33. Justice AC, McGinnis KA, Atkinson JH, Heaton RK, Young C, Sadek J, Madenwald T, Becker JT, Conigliaro J, Brown ST, Rimland D, Crystal S, & Simberkoff M, & Team, for the V. 5 P. (2004). Psychiatric and neurocognitive disorders among HIV-positive and negative veterans in care: veterans aging cohort five-site study. AIDS, 18(Suppl. 1), 49–59. 10.1097/00002030-200401001-00008 [DOI] [PubMed] [Google Scholar]
  34. Kemeny ME. (2003). An interdisciplinary research model to investigate psychosocial cofactors in disease: Application to HIV-1 pathogenesis. Brain, Behavior, and Immunity, 17(1), S62–S72. 10.1016/s0889-1591(02)00069-7 [DOI] [PubMed] [Google Scholar]
  35. Kongs SK, Thompson LL, Iverson GL, & Heaton RK. (2000). WCST-64: Wisconsin card sorting test-64 card version, professional manual. Psychological Assessment Resources. [Google Scholar]
  36. Leserman J, Petitto JM, Perkins DO, Folds JD, Golden RN, & Evans DL. (1997). Severe stress, depressive symptoms, and changes in lymphocyte subsets in human immunodeficiency virus—infected men. Archives of General Psychiatry, 54(3), 279–285. 10.1001/archpsyc.1997.01830150105015 [DOI] [PubMed] [Google Scholar]
  37. Liu B, & Li W. (2020). Association between education level and HIV-associated neurocognitive disorder: A systematic review and meta-analysis. Sexual Health, 17(6), 475. 10.1071/SH20057 [DOI] [PubMed] [Google Scholar]
  38. Lorant V, Deliège D, Eaton W, Robert A, Philippot P, & Ansseau M. (2003). Socioeconomic inequalities in depression: A meta-analysis. American Journal of Epidemiology, 157(2), 98–112. 10.1093/aje/kwf182 [DOI] [PubMed] [Google Scholar]
  39. Marquine MJ, Flores I, Kamat R, Johnson N, Umlauf A, Letendre S, Jeste D, Grant I, Moore D, & Heaton RK. (2018). A composite of multisystem injury and neurocognitive impairment in HIV infection: Association to everyday functioning. Journal of Neurovirology, 24(5), 549–556. 10.1007/s13365-018-0643-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Mayer JS, Bittner RA, Nikolić D, Bledowski C, Goebel R, & Linden DEJ. (2007). Common neural substrates for visual working memory and attention. NeuroImage, 36(2), 441–453. 10.1016/j.neuroimage.2007.03.007 [DOI] [PubMed] [Google Scholar]
  41. Miech RA, & Shanahan MJ. (2000). Socioeconomic status and depression over the life course. Journal of Health and Social Behavior, 41(2), 162–176. JSTOR. 10.2307/2676303 [DOI] [Google Scholar]
  42. Morgan EE, Iudicello JE, Weber E, Duarte NA, Riggs PK, Delano-Wood L, Ellis R, Grant I, & Woods SP. (2012). Synergistic effects of HIV infection and older age on daily functioning. Journal of Acquired Immune Deficiency Syndromes, 61(3), 341–348. 10.1097/QAI.0b013e31826bfc53 [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Nanni MG, Caruso R, Mitchell AJ, Meggiolaro E, & Grassi L. (2015). Depression in HIV infected patients: A review. Current Psychiatry Reports, 17(1), 530. 10.1007/s11920-014-0530-4 [DOI] [PubMed] [Google Scholar]
  44. Page-Shafer K, Delorenze N, Satatiano WA, & Winkelstein W. (1996). Comorbidity and survival in HIV-infected men in the San Francisco men’s health survey. Aep, 6 (5), 420–430. 10.1016/s1047-2797(96)00064-6 [DOI] [PubMed] [Google Scholar]
  45. Paolillo EW, Pasipanodya EC, Moore RC, Pence BW, Atkinson JH, Grelotti DJ, Grant I, Heaton RK, & Moore DJ. (2020). Cumulative burden of depression and neurocognitive decline among persons with HIV: A longitudinal study. JAIDS Journal of Acquired Immune Deficiency Syndromes, 84(3), 304–312. 10.1097/QAI.0000000000002346 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Pellowski JA, Kalichman SC, Matthews KA, & Adler N. (2013). A pandemic of the poor: Social disadvantage and the U.S. HIV epidemic. The American Psychologist, 68(4), 197–209. 10.1037/a0032694 [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. R Core Team. (2022). R: A language and environment for statistical## computing. R Foundation for Statistical Computing. https://www.R-project.org/ [Google Scholar]
  48. Rubin LH, & Maki PM. (2019). HIV, depression, and cognitive impairment in the era of effective antiretroviral therapy. Current HIV/AIDS Reports, 16(1), 82–95. 10.1007/s11904-019-00421-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Shimizu SM, Chow DC, Valcour V, Masaki K, Nakamoto B, Kallianpur KJ, & Shikuma C. (2011). The impact of depressive symptoms on neuropsychological performance tests in HIV-infected individuals: A study of the Hawaii aging with HIV Cohort. World Journal of AIDS, 1(4), 139–145. 10.4236/wja.2011.14020 [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Thames AD, Kuhn TP, Mahmood Z, Bilder RM, Williamson TJ, Singer EJ, & Arentoft A. (2018). Effects of social adversity and HIV on subcortical shape and neurocognitive function. Brain Imaging and Behavior, 12(1), 96–108. 10.1007/s11682-017-9676-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Unnikrishnan B, Jagannath V, Ramapuram JT, Achappa B, & Madi D. (2012). Study of depression and its associated factors among women living with HIV/AIDS in Coastal South India. ISRN AIDS, 2012, 1–4. 10.5402/2012/684972 [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Venables WN, & Ripley BD. (2002). Modern applied statistics with S (Fourth). Springer. https://www.stats.ox.ac.uk/pub/MASS4/ [Google Scholar]
  53. Wittchen HU, Robins LN, Cottler LB, Sartorius N, Burke JD, & Regier D. (1991). Cross-cultural feasibility, reliability and sources of variance of the Composite International Diagnostic Interview (CIDI). British Journal of Psychiatry, 159(5), 645–653. 10.1192/bjp.159.5.645 [DOI] [PubMed] [Google Scholar]
  54. Wood E, Montaner JS, Chan K, Tyndall MW, Schechter MT, Bangsberg D, O’Shaughnessy MV, & Hogg RS. (2002). Socioeconomic status, access to triple therapy, and survival from HIV-disease since 1996. AIDS, 16(15), 2065–2072. 10.1097/00002030-200210180-00012 [DOI] [PubMed] [Google Scholar]
  55. Woods SP, Rippeth JD, Frol AB, Levy JK, Ryan E, Soukup VM, Hinkin CH, Lazzaretto D, Cherner M, Marcotte TD, Gelman BB, Morgello S, Singer EJ, Grant I, & Heaton RK. (2004). Interrater reliability of clinical rating and neurocognitive diagnoses in HIV. Journal of Clinical and Experimental Neuropsychology, 26 (6), 759–778. 10.1080/13803390490509565 [DOI] [PubMed] [Google Scholar]
  56. World Health Organization. (2016). International statistical classification of diseases and related health problems (10th ed.). World Health Organization. https://icd.who.int/browse10/2016/en [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Zhu S, Zhao J, Chen Z, & Wang Y. (2018). Influential factors on cognitive performance in middle-aged cohort. Medicine, 97(37), e12033. 10.1097/MD.0000000000012033 [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES